1. Field of the Invention
The present invention relates to a method for manufacturing an exposure mask which is utilized in the manufacture of semiconductor integrated circuits and the like, and, in particular to a method for manufacturing an exposure mask which is provided with a phase-shifting mask by which the phase of the light is shifted.
2. Description of Background
In recent years, semiconductor circuits have become integrated on a large scale, so that a lithography technique is important in the manufacture of large-scale semiconductor integrated circuits.
Nowadays, light sources such as g-rays, i-rays, exicimer lasers, and X-rays have been adopted in manufacturing large-scale semiconductor integrated circuits by utilizing the lithography technique. Also, new resist materials which are chemically changed by exposed light generated by the light source have been developed to utilize the light source. Moreover, resist processing techniques such as a multi-layer resist method, a contrast enhanced lithography (CEL) method, and an image reverse method have been developed.
On the other hand, when the large-scale semiconductor integrated circuits are manufactured, an exposure mask is required to shield certain fields specified in the resist. That is, a striped pattern which is formed by alternately arranging dark and bright light regions is made by the light transmitted through the exposure mask. In detail, the specific fields in the resist aligned with the dark light region are not exposed by the light transmitted through the exposure mask so that the specific fields in the resist are selectively removed or remain after a prescribed removal process is implemented.
Therefore, the light intensity of the boundary between the bright and dark light regions must be considerably lowered to clarify the boundary, so that the resolution of the transmitted light can be improved.
Recently, a conventional method for manufacturing a exposure mask with a phase-shifting mask by utilizing a phase-shifting method has been proposed to improve the resolution of the transmitted light (lectured papers 8, page 497, lecture No. 4a-K-7 in the Applied Physics Society of Japan, autumn, 1988).
The phase-shifting method is described with reference to FIGS. 1A to 1F and 2 as follows.
FIGS. 1A, 1C, and 1E are cross sectional views showing a conventional method for manufacturing a conventional exposure mask which is utilized to make line sections and space sections on an exposed semiconductor device.
FIGS. 1B, 1D, and 1F are cross sectional views showing another conventional method for manufacturing another conventional exposure mask which is utilized to make an isolated space section on an exposed semiconductor device.
As shown in FIGS. 1A, 1B, a shading layer 1 made of chromium (Cr) or chromium oxide (Cr.sub.2 O.sub.3) is deposited to a thickness of about 1000 angstroms (.ANG.) on a quartz substrate 2 by a sputtering method. The exposure light cannot pass through the shading layer 1, but can pass through the quartz substrate 2 without a substantial decrease in intensity.
Thereafter, a resist layer 3 for protecting the shading layer 1 from being etched is coated and patterned on the shading layer 1 by an electron beam unit. Thereafter, as shown in FIGS. 1C, 1D, the shading layer 1 not covered by the resist layer 3 is etched by a wet etching method or a reactive ion etching method, and the shading layer 1 covered by the resist layer 3 is protected from the etching process. Therefore, the shading layer i not etched by the etching method is patterned to form a patterned shading layer 4 before the patterned resist layers 3 are removed. Openings 5 between the patterned shading layers 4 are wider than a pair of auxiliary openings 6 between the patterned shading layers 4.
Thereafter, as shown in FIG. 1E, the openings 5 are alternately deposited by a plurality of phase shifters 7 using a chemical vapour depositing (CVD) method. Moreover, as shown in FIG. 1F, each auxiliary opening 6 is covered by a phase shifter 8 by the CVD method, while the opening 5 is not covered by a phase shifter. The height of the phase shifters 7, 8 is a prescribed value H=.lambda./2(n-1). Here, the symbol .lambda. indicates a wavelength of an exposure light transmitted from the quartz substrate 2 to the phase shifter 7 (8) or the opening 5, and the symbol n indicates the refractive index of the phase shifters 7, 8. Therefore, the phase of the exposure light transmitted through the quartz substrate 2 and the phase shifter 7, or 8 is shifted by a half wavelength .lambda./2 as compared with the exposure light transmitted the opening 5 through the quartz substrate 2.
FIGS. 2A, 2B are cross sectional views of the conventional exposure masks manufactured by the method shown in FIGS. 1A to 1F. FIGS. 2C, 2D show the amplitude of the exposure light transmitted from the quartz substrate 2 to the phase shifter 7 (8) or the opening 5.
As shown in FIGS. 2A, 2B, the phase of the exposure light transmitted through the phase shifter 7 (8) is shifted by a half wavelength as compared with the exposure light transmitted through the opening 5. Moreover, the exposure light transmitted through the phase shifter 7 (8) or the opening 5 is optically diffracted by the patterned shading layers 4, so that the various exposure lights transmitted through the exposure mask optically interfere with one another.
Therefore, as shown in FIG. 2C. 2D, the diffracted lights of which the amplitudes are shown by broken lines are superimposed on one another so that an interfered light of which the amplitude is shown by a solid line is formed. That is, the intensity of the exposure light is considerably decreased in optical paths specified between the phase-shifted light transmitted through the phase shifter 7 (8) and the exposure light transmitted through the opening 5.
Accordingly, because the light intensity is considerably decreased in the specific optical paths, the difference between dark and bright light regions alternately arranged by the light transmitted through the exposure mask can be clearly distinguished.
For example, in the case of the conventional exposure mask which is utilized to make line sections and space sections on a negative type of resist layer of an exposed semiconductor device, as shown in FIGS. 2A, 2C, the line sections are made in positions aligned with the opening 5 or the phase shifter 7 by exposure to the bright light transmitted through the exposure mask. Moreover, the space sections are made in positions aligned with the shading layer 1 by exposure to the dark light transmitted through the exposure mask.
On the other hand, in the case of a conventional exposure mask which is utilized to make an isolated space section on a positive type of resist layer of an exposed semiconductor device, as shown in FIGS. 2B, 2D, the intensity of the exposure light transmitted through the phase shifter 8 is, as is well known, considerably smaller than that of the exposure light transmitted through the opening 5 because the width of the auxiliary opening 6 is narrower than the width of the opening 5. Therefore, the shifted exposure light transmitted through the phase shifter 8 is considerably decreased after the shifted light is subjected to optical interference from the exposure light transmitted through the opening 5. As a result, the resist layer on the exposed semiconductor device is not exposed by the shifted exposure light because of insufficient light. That is, an isolated space section is made on the resist layer aligned with the opening 5 by exposing the bright light transmitted through the exposure mask, while the resist layer aligned with the phase shifters 8 is not exposed.
However, many drawbacks are generated in the above the exposure mask and the above method for manufacturing the exposure mask.
A first drawback of the method is as follows.
The phase shifters 7 must be alternately arranged in the openings 5 between the patterned shading layers 4 in cases where the line sections and the space sections are formed on the semiconductor device. Therefore, a manufacturing method includes at least a first pattern forming step for forming the patterned shading layers 4, a second pattern forming step for alternately forming the phase shifters 7, and an alignment step for aligning the phase shifters 7 with the patterned shading layers 4. Usually, an alignment function is not provided in the electron beam exposure unit for patterning the shading layer, so that an electron beam exposure unit provided with this alignment function must be developed.
Therefore, it is difficult to implement the above conventional manufacturing method because a large scale unit is required and the cost is very high.
A second drawback of the method is as follows.
A large amount of mask pattern data and lithography data for alternately arranging the phase shifters is required. Therefore, a large amount of complicated data processing is required to manufacture the above exposure mask.
A third drawback of the method is as follows.
The auxiliary openings 6 must be smaller than the opening 5 to decrease the light intensity. Therefore, in cases where the isolated space section must be the minimum size, the auxiliary openings 6 must be smaller than the minimum size.
A first drawback of the exposure mask is as follows.
Because the phase shifters 7,8 are made by the CVD method, the refractive index of the shifters 7, 8 is not exactly the same as that of the substrate 2 regardless of whether the material of the shifters 7, 8 is the same as that of the substrate 2. Therefore, a multi-reflection is generated between the substrate 2 and the phase shifter 7 (8) so that the intensity of the exposure light transmitted through the phase shifter 7 (8) is smaller than that of the exposure light transmitted through the opening 5 in cases where the line and space sections are formed on the semiconductor device. As a result, the intensity of the dark light passing through the specific paths aligned with the shading layers 4 is not fully decreased.
A second drawback of the exposure mask is as follows.
When the exposure mask is washed by an acid processing solution, the phase shifters 7, 8 are eroded away by the solution because the shifters 7, 8 are not very dense.
Next, three types of conventional exposure mask with a phase-shifting mask, such as a Levenson type, a self-aligned type, and a shift edge type are described.
FIG. 3A is a cross sectional view of a Levenson type of exposure mask with a phase-shifting mask. FIG. 3B shows the amplitude distribution of an exposure light transmitted through the mask shown in FIG. 3A in cases where the light is regarded to be not optically diffracted. FIG. 3C shows the amplitude distribution of an exposure light after the light is optically diffracted by the shading layers shown in FIG. 3A. FIG. 3D shows the intensity distribution of the exposure light of which the amplitude is shown by a solid line in FIG. 3C. FIG. 3E shows space and line sections made on a semiconductor device after the device is exposed by the exposure light transmitted through the exposure mask shown in FIG. 3A.
The Levenson type of exposure mask with a phase-shifting mask shown in FIG. 3A comprises a quartz substrate 11 for transmitting an exposure light, shading layers 12 arranged on the quartz substrate 11 at regular intervals for shading the exposure light transmitted through the substrate 11, phase shifters 13 alternately arranged between the shading layers 12 for shifting the phase of the exposure light transmitted through the substrate 11 by a half wavelength as compared with the phase of the light transmitted through openings 14 alternately arranged between the shading layers 12.
A first exposure light transmitted through both the substrate 11 and the opening 14 is diffracted by the shading layers 12. Therefore, the rectangular amplitude distribution of the first exposure light shown in FIG. 3B is changed to the waveform distribution shown by the upper broken line in FIG. 3C.
On the other hand, the phase of a second exposure light transmitted through both the substrate 11 and the phase shifter 13 is shifted by a half wavelength as compared with the first exposure light transmitted through the opening 14. Simultaneously, the shifted second exposure light is diffracted by the shading layers 12. Therefore, the rectangular amplitude distribution of the second exposure light shown in FIG. 3B is changed to the waveform distribution as shown by the lower broken line in FIG. 3C.
Therefore, these exposure lights are subjected to optical interference with one another so that the amplitude of the interfered light is changed to a superposed pattern shown by the solid line in FIG. 3D.
Accordingly, the light intensity is considerably decreased in an optical path aligned with the shading layers 12 so that the space and line sections can be formed on the semiconductor as shown in FIG. 3E.
The Levenson type of exposure mask is easily manufactured and is utilized to manufacture a semiconductor device provided with line sections and space sections because the phase shifters 13 are alternately arranged between the shading layers 12. Therefore, the exposure mask is not appropriate for manufacturing a semiconductor device provided with the isolated space section.
FIG. 4A is a cross sectional view of a self-aligned type of exposure mask with a phase-shifting mask. FIG. 4B shows the amplitude distribution of an exposure light transmitted through the mask shown in FIG. 4A in cases where the light is regarded to be not optically diffracted. FIG. 4C shows an amplitude distribution of an exposure light after the light is optically diffracted by phase shifters shown in FIG. 4A. FIG. 4D shows an intensity distribution of the exposure light of which the amplitude is shown in FIG. 4C. FIG. 4E shows space and line sections made on a semiconductor device after the device is exposed by the exposure light transmitted through the exposure mask shown in FIG. 4A.
A self-aligned type of exposure mask with a phase-shifting mask shown in FIG. 4A comprises a quartz substrate 21 for transmitting an exposure light, shading layers 22 arranged on the quartz substrate 21 at regular intervals for shading the exposure light transmitted through the substrate 21, phase shifters 23 arranged and overhung by a regular width .delta.1 on the shading layers 22 for shifting the phase of the exposure light transmitted through the substrate 21 by a half wavelength of the light as compared with the phase of the light transmitted through openings 24 between the overhanging parts of the phase shifters 23.
A first exposure light transmitted through both the substrate 21 and the opening 24 is diffracted by the overhanging parts of the phase shifters 23. Therefore, the rectangular amplitude distribution of the light shown in FIG. 4B is changed to the wave form distribution shown by the upper broken line in FIG. 4C.
On the other hand, the phase of a second exposure light transmitted through both the substrate 21 and the overhanging part of the phase shifter 23 is shifted by a half wavelength as compared with the first exposure light transmitted through the opening 24. Simultaneously, the shifted second exposure light is diffracted by the phase shifter 23. Therefore, the rectangular amplitude distribution of the second exposure light shown in FIG. 4B is changed to a waveform distribution shown by the lower broken line in FIG. 3C. In this case, because the overhanging part of the phase shifter 23 transmitting the second exposure light is a narrow width 81 such as 0.04 .mu.m, the intensity of the second exposure light is, as is well known, decreased.
Therefore, when these exposure lights are subjected to optical interference with one another in an optical path aligned with the overhanging part of the phase shifter 23, the intensity of the interfered light superimposed on these lights in the optical path is considerably decreased as shown by the solid line in FIG. 4C so that the amplitude of the interfered light is changed to the superimposed pattern shown by the solid line in FIG. 4D.
Accordingly, the light intensity is considerably decreased in an optical path aligned with the phase shifters 23 so that the space and line sections or the isolated space section can be formed on the semiconductor as shown in FIG. 4E.
The self-aligned type of exposure mask is appropriate for manufacturing a semiconductor device provided with the line and space sections or the isolated space section because the phase shifters 23 are arranged on all the shading layers 22, and the interval between the phase shifters 23 can be arbitrarily set.
Therefore, the exposure mask can be utilized in manufacturing any type of semiconductor circuit.
However, the manufacturing steps are complicated because the phase shifters 23 must overhang by the regular width .delta.1 on the shading layers 22.
FIG. 5A is a cross sectional view of a shift edge type of exposure mask with a phase-shifting mask. FIG. 5B shows the amplitude distribution of an exposure light transmitted through the mask shown in FIG. 5A in cases where the light is regarded to be not optically diffracted. FIG. 5C shows the amplitude distribution of an exposure light after the light is optically diffracted by the phase shifter as shown in FIG. 5A. FIG. 5D shows the intensity distribution of the exposure light for which the amplitude is shown in FIG. 5C. FIG. 5E shows space and line sections made on a semiconductor device after the device is exposed by the exposure light transmitted through the exposure mask shown in FIG. 5A.
A shift edge type of exposure mask with a phase-shifting mask shown in FIG. 5A comprises a quartz substrate 31 for transmitting an exposure light, phase shifters 32 arranged at regular intervals on the quartz substrate 31 for shifting the phase of the exposure light transmitted through the substrate 31 by a half wavelength as compared with the phase of the light transmitted through the openings 33 between the phase shifters 32.
A first exposure light transmitted through both the substrate 31 and the opening 33 is diffracted by the phase shifters 32. Therefore, the rectangular amplitude distribution of the light shown in FIG. 5B is changed to the waveform distribution shown by the upper broken line in FIG. 5C.
On the other hand, the phase of a second exposure light transmitted through both the substrate 31 and the phase shifter 32 is shifted by a half wavelength as compared with the first exposure light transmitted through the opening 33. Simultaneously, the shifted second exposure light is diffracted by the phase shifter 32. Therefore, the rectangular amplitude distribution of the second exposure light shown in FIG. 5B is changed to the waveform distribution shown by the lower broken line in FIG. 5C.
Therefore, when these exposure lights are subjected to optical interference with one another in an optical path aligned with a narrow boundary region between the phase shifter 32 and the opening 33, the intensity of the interfered light superimposed on these lights in the optical path is considerably decreased as shown by the solid line in FIG. 5C so that the amplitude of the interfered light is changed to the superimposed pattern shown by the solid line in FIG. 5D.
Accordingly, a resist layer on a semiconductor device aligned with the narrow boundary is underexposed, so that narrow space sections can be formed on the semiconductor device as shown in FIG. 5E.
The shift edge type of exposure mask can be easily manufactured because no shading layer is provided for the exposure mask.
However, an arranged pattern of the phase shifters 32 differs from a formed pattern of the space sections on the semiconductor device because the space section is produced in specific fields aligned with the narrow boundary. Therefore, the exposure mask can be utilized to manufacture a specific type of semiconductor device.
FIG. 6 is a graphic view showing the resolving power of three types of exposure masks shown in FIGS. 3A, 4A, and 5A, this resolving power being estimated by utilizing an image contrast.
FIG. 7 is a graphic view showing the focus margin of the three types of exposure masks shown in FIGS. 3A, 4A, and 5A, the focus margin being expressed by an image contrast.
The resolving power and the focus margin are estimated using a computer. The resolving power is expressed by the minimum pattern size in FIG. 6. The focus margin is expressed by allowable de-focus values.
As shown in FIG. 6, the minimum pattern size for manufacturing a semiconductor circuit can be estimated by utilizing an image contrast between the dark and bright lights composing the light transmitted through the exposure mask. That is, the minimum pattern size is determined, when the value of the image contrast agrees with an allowable image contrast value.
For example, the minimum pattern size Is estimated to be 0.17 .mu.m for the Levenson and shift edge types of exposure masks (.sigma.=0.2), 0.22 .mu.m for the Levenson type of exposure mask (.sigma.=0.5), 0.27 .mu.m for the self-aligned type of exposure mask (.sigma.1=0.04 .mu.m, .sigma.=0.5), and 0.3 .mu.m for an exposure mask with no phase-shifting mask on condition that a numerical aperture NA equals 0.42 and the wavelength .lambda. of the exposure light equals 248 nm. Here, as is well known, the symbol .sigma. indicates a ratio of the NA of an imaging lens to the NA of a focusing lens.
As shown in FIG. 7, de-focus values can be calculated by utilizing the value of the image contrast estimated in cases where the semiconductor device exposed by an exposure light is arranged out of focus. That is, when the value of the image contrast agrees with the allowable image contrast value, the absolute de-focus value indicates the focus margin.
For example, the focus margin is estimated to be 0.9 .mu.m in the Levenson and shift edge types of exposure masks (.sigma.=0.2), 0.45 .mu.m in the Levenson type of exposure mask (.sigma.=0.5), and 0.35 .mu.m in the self-aligned type of exposure mask (.sigma.1=0.04 .mu.m, .sigma.=0.5), while no focus margin is estimated for the exposure mask with no phase-shifting mask on condition that the numerical aperture NA equals 0.42 and the wavelength .lambda. of the exposure light equals 248 nm.
Therefore, the resolving power and the focus margin are improved to some extent by utilizing the Levenson and shift edge types of exposure masks as compared with the conventional method. Moreover, the resolving power and the focus margin are slightly improved by utilizing the shift edge type of exposure mask.
However, the Levenson type of exposure mask is not appropriate for use in practice because the Levenson type requires a specific arrangement whereby the phase shifters 13 and shading films 12 are repeatedly arranged at regular intervals.
Moreover, all types of exposure masks are made by a chemical vapour deposition (CVD) method in which vapourized silicone oxide (SiO.sub.2) is chemically deposited on a quartz substrate to form phase shifters. Therefore, the phase shifters are rapidly formed so that it is difficult to control their thickness. Moreover, it is difficult to uniformly adjust the thickness of all phase shifters because the deposition rate is not uniform over the entire quartz substrate in the CVD method. According to the principle of the phase-shifting method, the unevenness in the thickness of the phase shifters leads to unevenness in the phase of the shifted exposure light, so that the resolving power and the focus margin deteriorate to a considerable extent.
Further, the refractive index of the phase shifter differs from that of the quartz substrate regardless of whether the same material is utilized. Therefore, multi-reflection of exposure light is generated between the quartz substrate and the phase shifter.